Loading ...
Sorry, an error occurred while loading the content.

GEOSTATS: 2nd Int. Symp. on Sensitivity Analysis of Model Output

Expand Messages
  • Gregoire Dubois
    Second International Symposium on Sensitivity Analysis of Model Output (SAMO 98) Venice, Ca Dolfin, April 19-22 1998 ***************** Final Call for Papers
    Message 1 of 1 , Jul 3, 1997
      Second International Symposium on
      Sensitivity Analysis of Model Output

      (SAMO 98)

      Venice, Ca' Dolfin, April 19-22 1998

      Final Call for Papers

      A special issue of the journal Computer Physics Communications (CPC),
      published by Elsevier Science, will be devoted to SAMO98. About 30
      contributed papers presented at the conference will be selected to appear
      in this special issue. All CPC submissions are subject to peer review.

      Sensitivity analysis (SA) aims to ascertain how a given model (numerical
      or otherwise) depends on its input parameters. SA can be thought of as a
      clearly defined and delimited discipline. Yet it occurs in a variety of
      different settings and fields, and it is hard to keep track of its progress.
      In spite of its apparent simplicity, in practice sensitivity analysis may be
      an arduous task even for models of limited complexity. Different branches of
      Mathematics and Statistics have produced very different and original ways
      to carry out an efficient SA. The sensitivity analysis community -
      composed of investigators from a wide spectrum of disciplines ranging
      from Mechanics to Economics - met for the first time in September 1995
      (see http://www.ei.jrc.it/samo). This second International symposium will
      build on the first successful cross-fertilization exercise to strengthen
      the role of sensitivity analysis as a prerequisite for model building in all
      settings - be it diagnostic or prognostic - and all fields - be it
      mathematical, natural or social sciences - where models are used and
      developed (see background).

      Where and When
      SAMO 98 will be held in the "Aula Magna di Ca' Dolfin", University of
      Venice, Venice, Italy. The conference will start from 2:30pm on Sunday 19
      April and will last till 1:30pm on Wednesday 22 April 1998. Registration will
      start at 11am on Sunday, 19 April 1998.

      Scientific Programme
      The provisional outline of the conference programme is as follows:

      Sensitivity Analysis and Epistemology
      Variance-based Methods
      Screening Methods
      Bayesian Sensitivity Analysis
      Inverse Problems
      Applications of SA

      Invited Speakers Programme
      Invited lectures are given by members of the SAMO Scientific Committee
      and guest speakers(*). Provisional titles of the invited lectures are as

      Terry Andres, AECL Whiteshell Labs, Canada
      SAMPLE: Software to Generate Experimental Designs for Large Sensitivity
      Analysis Experiments

      Bruce Beck, University of Georgia, USA
      Structural Error and the Identification of Structural Change in the
      Behaviour of Environmental Systems

      Russell Cheng*, University of Kent, UK
      Model Selection by Parametric Bootstrapping: A Powerful Alternative to Bayes

      Roger Cooke, Delft University of Technology, Netherlands
      Active Issues in Recent Applications of Uncertainty Analysis

      David Draper, University of Bath, UK
      The Bayesian Approach to Model Uncertainty: Methods and Applications

      David Hamby*, University of Michigan, USA
      Practical Applications of Sensitivity Analysis in Environmental Modelling

      Jon Helton, University of Arizona, USA
      Uncertainty and Sensitivity Analysis in Performance Assessment for the
      Waste Isolation Pilot Plant

      Eduard Hoefer, Gesellshaft fuer Anlagen and Reaktorsicherheit (GRS), Germany
      Sensitivity Analysis for Long-Running Computer Models

      Toshimitsu Homma, Japan Atomic Energy Research Institute, Japan
      Uncertainty and Sensitivity Analysis in Accident Consequence Assessments

      David Insua, Fabrizio, Ruggeri, Jacinto Martin, Politechnic University
      of Madrid, Spain
      Bayesian Sensitivity Analysis: some recent developments

      Jack Kleijnen, Tilburg University, Netherlands
      Statistical Theory on Design of Experiments, Applied to Simulation

      Masato Koda, IBM Research, Japan
      Stochastic Sensitivity Analysis and Adaptive Critic Reinforcement Learning

      Michael McKay, Los Alamos National Laboratory, USA
      Evaluating Uncertainty in Stochastic Simulation Models

      Naomi Oreskes*, New York University, USA
      Informing Judgement: Models, Sensitivity Analysis, and the Real World

      Hershel Rabitz, Princeton University, USA
      Managing the Tyranny of Parameters in Mathematical Modeling

      Andrea Saltelli, Environment Institute, JRC Ispra, Italy
      Are Variance-based SA Methods Healthy for Your Models?

      Marian Scott, University of Glasgow, UK
      Sources of Uncertainty in Environmental Models and their Quantification

      David Stanners, European Environment Agency, Denmark
      The Need for Sensitivity Analysis in Integrated Environmental Assessment

      Tamas Turanyi, Central Research Institute for Chemistry, Hungary
      Development of Chemical Kinetic Models using Sensitivity Analysis Information

      In addition to the above and contributed paper sessions, there will be
      tutorial and poster sessions. Provisional arrangement for these sessions are
      as follows:

      Tutorials --- Sunday, 19 April, 2:30pm - 5pm

      Posters --- Monday & Tuesday, 20 & 21 April, 4pm - 6:30pm

      Contributed papers will be presented in oral (20 minutes + 5 minutes
      discussion) or posters (board area is 150cm by 100cm) sessions.

      The working language of the symposium is English.

      SAMO 98 is organised jointly by the Environment Institute of the Joint
      Research Centre of the EC and the University of Venice, Italy.

      SAMO Scientific Committee:

      Terry Andres, Atomic Energy of Canada Ltd. (AECL)
      George Apostolakis, Massachusetts Institute of Technology , (US)
      Bruce Beck, The University of Georgia (US)
      Roger Cooke, Technical University of Delft, (NL)
      David Draper, University of Bath, (UK)
      Jon Helton, University of Tempe , Arizona (US)
      Eduard Hoefer, GRS, (D)
      Toshimitsu Homma, Japan Atomic Energy Research Institute (JAERI)
      Dave Rios Insua, Politechnic University of Madrid, (E)
      Jack Kleijnen, Tilburg University, (NL)
      Masato Koda, IBM Research, Tokyo, (J)
      Michael McKay, Los Alamos National Laboratory NM (US)
      Roberto Pastres, University of Venice, (I)
      Hershel Rabitz, Princeton University, NJ (US)
      Andrea Saltelli, Environment Institute, JRC Ispra, (I)
      Marian Scott, University of Glasgow, (UK)
      David Stanners, European Environment Agency, (DK)
      Ilya Sobol', Russian Academy of Science, Moscow (CIS)
      Tamas Turanyi, Central Research Institute for Chemistry, Budapest, (H)
      Henning Von Maravic', European Commission DG XII, Bruxelles (B)
      Peter Young, University of Lancaster (UK)

      SAMO 98 Organizing Committee:

      Environment Institute, Joint Research Centre, Ispra -

      Andrea Saltelli (andrea.saltelli@...)
      Karen Chan (karen.chan@...)
      Stefano Tarantola (stefano.tarantola@...)

      Department of Physical Chemistry, University of Venice -

      Roberto Pastres (pastres@...)

      JRC Public Relations Unit, Ispra -

      Dorit Schlittenhardt (dorit.schlittenhardt@...)

      Abstracts and Pre-proceedings
      A three page abstract, including a title, author name(s) and address(es),
      should be submitted to Dr Karen Chan, Environment Institute, European
      Commission, Joint Research Centre, TP 272, 21020 Ispra (VA), ITALY,
      no later than 30 June 1997.

      It is possible to send the abstracts by electronic mail (to
      "karen.chan@...") and via the Web. In this case, documents should be
      prepared in LATEX, WORD for Windows, or WORDPERFECT format.
      Instructions for Web submission are available at "http://www.ei.jrc.it/samo".
      *** It is also possible to submit your abstract at the EI ftp site, please
      ask the organisers for further details. ***

      Authors will receive notification of acceptance of their abstract by 15
      September 1997. Final versions of the abstracts for oral and poster
      presentation must be submitted by 21 December 1997 to ensure
      inclusion in the pre-proceedings. The pre-proceedings will be published
      by the European Commission.

      A special issue of the journal Computer Physics Communications (CPC),
      published by Elsevier Science, will be devoted to SAMO98. About 30
      contributed papers presented at the conference will be selected to appear
      in this special issue. All CPC submissions are subject to peer review.

      Important Dates
      Submission of Abstracts (3 pages): June 30, 1997

      Notification of Acceptance to Authors: September 15, 1997

      Registration: October 15, 1997

      Final manuscripts: December 21, 1997

      Registration Fee
      The conference fee per participant, inclusive of the pre-proceedings and
      conference dinner, is

      Lit. 440,000 registration before 15 October 1997

      Lit. 540,000 registration after 15 October 1997

      Students will receive a Lit. 200,000 discount. Proof of status is required
      and should be submitted with your registration form. Special allowances will
      be given to scientists who come from Eastern European countries, please ask
      the organisers for further details.

      Payment may be made by

      Bank transfer to Banca Nazionale del Lavoro (BNL) di Varese, c/o
      CCR Ispra (Italy), Account No.: 01005-50340-5500/SAMO98
      (Please quote "SAMO 98" in all communications.), or
      International bank order/draft made payable to "European Commission, JRC

      Please send your payment with your registration form to Dorit
      Schlittenhardt, Public Relations Unit, European Commission Joint
      Research Centre, TP 020, 21020 Ispra (VA), ITALY.

      Please make sure that

      * the payment is in Lira, otherwise add 10% of the registration fee
      * to the total amount;
      * all transfer fees must be met by the applicants;
      * a copy of the bank transfer is enclosed with your registration form.

      Access to Venice
      Venice can be reached by road, by train, by plane and by sea!
      The train reaches the historical city at the station of "Venezia S.L" which is
      located in front of the Grand Canal.
      Venice can be accessed by car, but parking could be difficult and expensive.
      The car terminal of "Piazzale Roma" is near the Congress room.
      Venice airport is connected with the car terminal of "Piazzale Roma" by public
      and private regular bus services. Buses take about 20 minutes to
      reach "Piazzale Roma".
      Venice is connected with Mestre by train and bus services; the journey takes
      approximately 15 minutes.

      General Information
      Contact address:

      SAMO 98
      c/o Environment Institute (EI)
      European Commission Joint Research Centre
      TP 272, 21020 Ispra (VA), ITALY
      Fax. no.: ++ 39 332 78 5733
      Email: karen.chan@...

      For information about Venice, please contact:

      Dr Roberto Pastres
      Department of Physical Chemistry
      University of Venice
      Dorsiduro 2137, Venice, ITALY
      Fax. no.: ++ 39 415 29 8594
      Email: pastres@...

      >From "The role of sensitivity analysis in the corroboration of models and
      its links to model structural and parametric uncertainty" by A. Saltelli and
      M. Scott, To appear on Rel. Eng. & Syst. Safety, Special Issue on
      sensitivity analysis.

      [...] a tentative condensed list of reasons why (and instances where) SA
      should be considered.

      (1) In control problems, SA may help to identify critical regions in
      the space of the input parameters.

      (2) In screening exercises, SA may help to locate a few influential
      parameters in systems with hundreds of uncertain inputs.

      (3) Variance based SA techniques are useful to ascertain if a subset of
      input parameters may account for (most of) the output variance.

      (4) Point (3) above may be used for mechanism reduction (dropping or
      fixing non relevant parts of the model) and for model lumping
      (building/extracting a model from a more complex one). See also the
      problem of model "relevance": are the parameters in the model input set
      relevant to the task of the model?

      (5) Point (3) above may also be used for model identification, by
      pinpointing the experimental conditions for which your ability to
      discriminate among model is maximum.

      (6) As in (5) above SA may be used for model calibration, to ascertain if
      the experiments with its related uncertainties will allow parameter
      estimation. Useful especially against ill conditioned problems.

      (7) SA may be coupled to optimization / search algorithms; by identifying
      the most important parameters SA may allow the dimensionality of the
      space where the search is made to be reduced.

      (8) As a quality assurance tool, to make sure that the dependence of the
      output from the input parameters in the model has a physical
      resemblance and explanation.

      (9) To solve an inverse problem, ie as a tool to extract parameters
      embedded into models, whose output does not correlate easily with the
      unknown input (eg in chemical kinetics, to extract kinetic constants of
      complex systems from measured yield rate of components; see also
      functional sensitivities).

      (10) To optimally allocate resources in R&D, by showing where it is more
      worthwhile to invest in order to reduce the model's range of uncertainty

      (11) To ascertain on a quantitative basis what fraction of my prediction
      uncertainty is due to parametric (or aleatory) uncertainty and how much to
      structural (or epistemic) uncertainty.

      (12) More specific applications exists such as SA based training of neural
      networks (instead of the back-propagation algorithm), SA based system to
      couple atmospheric transport of gaseous species with their chemical
      reaction, etc.

      Further Information
      Extensive background material on sensitivity analysis can be found at

      <Moderator of the ai-geostats mailing list>
      *To post a message to the list, send it to ai-geostats@....
      *As a general service to list users, please remember to post a summary
      of any useful responses to your questions.
      *To unsubscribe, send email to majordomo@... with no subject and
      "unsubscribe ai-geostats" in the message body.
      DO NOT SEND Subscribe/Unsubscribe requests to the list!
    Your message has been successfully submitted and would be delivered to recipients shortly.